Morgan T. Sammons, Norman B. Schmidt Combined Treatments for Mental Disorders_ a Guide to...

COMBINED TREATMENTS FOR MENTAL DISORDERS

A GUIDE TO PSYCHOLOGICAL AND PHARMACOLOGICAL INTERVENTION s

EDITED BY

MORGAN T. S M O N S

AND NORMAN B. SCHMIDT

American Psychological Association Washington, DC

Contents

Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii

Introduction: Toward a Psychological Model of Pharmacological Service Provision ............................ 3 Morgan 71 Sammons and Norman B. Schmidt

1. Combined Treatments for Mental Disorders: Clinical Dilemmas. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Morgan T Sammons

2. Prescriptive Authority for Psychologists: Law, Ethics,

Patrick H. DeLeon, Sharon E. Robinson Kurpius, and John L. Sexton

and Public Policy.. ...................................... 33

3. Comparative and Combined Treatments for

Martin M. Antony and Richard I? Swinson Obsessive-Compulsive Disorder ......................... 53

4. Combined Treatments for Phobic Anxiety Disorders . . . . . . 81 Norman B. Schmidt, Margaret Koselka, and Kelly Woolaway-Bickel

5. Combined Treatments of Insomnia. ...................... 111 Charles M. Morin

6. Combined Treatments for Depression .................... 131 Jeremy W. Pettit, Zachary R. Voelz, and Thomas E. Joiner, Jr

7. Combined Treatments and Rehabilitation of Schizophrenia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 William D. Spaulding, Dale L. Johnson, and Robert D. Coursey

8. Combined Treatments for Smoking Cessation . . . . . . . . . . . . 19 1 Marc E. Mooney and Dorothy K. Hatsukami

9. Combined Treatments for Substance Dependence. . . . . . . . . 215 Kathleen M, Carroll

10. Pharmacological and Psychological Treatments of Obesity and Binge Eating Disorder .............................. 239 Carlos M. Grilo

V

vi CONTENTS

11. Clinical Outcomes Assessment for the Practicing Clinician . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271 James M. Meredith, Michael J. Lambert, and John I? Drozd

. . .

Appendix: Generic and Trade Names of Drugs Cited in This Volume.. ............................................ 301

Glossary of Technical Terms ....................................... 307

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313

Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337

About the Editors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345

Contributors

Martin M. Antony, PhD, Anxiety Treatment and Research Centre, St. Joseph's Hospital, and Department of Psychiatry and Behavioral Neurosciences, McMaster University, Hamilton, Ontario, Canada

New Haven, CT

Maryland, College Park, MD

Senate, Washington, DC

Center, Peterson Air Force Base, CO

New Haven, CT

Psychiatry, University of Minnesota, Minneapolis, MN

Houston, Houston, TX

University, Tallahassee, FL

Psychology, Uniformed Services University of the Health Sciences, Bethesda, MD

Arizona State University, Tempe, AZ

Salt Lake City, UT

US Air Force, and PACAF Psychology Consultant, Hickam Air Force Base, HI

Program, University of Minnesota, Minneapolis, MN

Quebec, Canada


Clinic, Annapolis, MD

University, Columbus, OH

San Diego, CA

Nebraska at Lincoln, NE

Kathleen M. Carroll, PhD, Yale University School of Medicine,

Robert D. Coursey, PhD, Psychology Department, University of

Patrick H. DeLeon, PhD, JD, Office of Senator Daniel Inouye, U S .

John F. Drozd, PhD, Capt, USAF, BSC, 10th Medical Group, Life Skills

Carlos M. Grilo, PhD, Yale University School of Medicine,

Dorothy K. Hatsukami, PhD, Division of Neuroscience Research in

Dale L. Johnson, PhD, Department of Psychology, University of

Thomas E. Joiner, Jr., PhD, Department of Psychology, Florida State

Margaret Koselka, PhD, Department of Medical and Clinical

Sharon E. Robinson Kurpius, PhD, Counseling Psychology Program,

Michael J. Lambert, PhD, Brigham Young University,

James M. Meredith, Lt. Col., PhD, Prescribing Psychologist,

Marc E. Mooney, MA, Clinical Science and Psychopathology Research

Charles M. Morin, PhD, School of Psychology, Lava1 University,

Jeremy W. Pettit, MS, Department of Psychology, Florida State

Morgan T. Sammons, PhD, Mental Health Department, Naval Medical

Norman B. Schmidt, PhD, Department of Psychology, The Ohio State

John L. Sexton, PhD, Prescribing Psychologist, Naval Medical Center,

William D. Spaulding, PhD, Department of Psychology, University of

vii

viii CONTRIBUTORS

Richard P. Swinson, MD, Department of Psychiatry and Behavioral Neurosciences, Faculty of Health Sciences, McMaster University, Hamilton, Ontario, Canada


University, Columbus, OH

Zachary R. Voelz, BA, Department of Psychology, Florida State

Kelly Woolaway-Bickel, MA, Department of Psychology, The Ohio State

COMBINED TREATMENTS FOR MENTAL DISORDERS

Introduction: Toward a Psychological Model of Pharmacological

Service Provision

Morgan I: Sammons and Norman B. Schmidt

This book is aimed at psychologists and other mental health practitioners who desire to understand how psychotropic drugs can be combined with psychotherapy and other behavioral treatments to produce optimum patient outcome. Readers will discover that the science underlying combined treatments remains underdeveloped. This is in part a reflection of the inattention paid to investigating combined treatments, in part a reflection of guild-based biases that champion one form of treatment over another, and in part because of the complexity and increased costs associated with combined-treatment research designs.

As a number of chapters in this book attest, combined treatments may not represent the best option for many patients. In particular, the literature suggests that many anxiety disorders may be better treated with behavioral rather than pharmacological interventions. Behavioral treatments for phobic and other anxiety disorders are often more durable than are drug treatments, and they do not carry the risks of dependence that accompany the use of some pharmacological interventions for these disorders (the benzodiazepines). Nevertheless, not all patients are amenable to nondrug treatments because of choice, chronicity, or severity of condition. All of these factors might mitigate toward the addition of pharmacotherapy as an adjunct to behavioral treatment. It is therefore incumbent on the clinician to keep an open mind and not reject a treatment modality categorically. Clinicians who rely exclusively on psychotherapy commit as great an error as those who rely exclusively on pharmacology, for neither approach is likely to completely address the needs of all those who seek help. Flexibility in thinking and attention to the needs of the patient are far better guideposts to successful intervention than is reliance on drug company literature or the opinions of therapists who dogmatically reject all but psychotherapy.

This book will assist clinicians in understanding the research literature on combined treatments. To the extent that the literature allows, algorithms or specific treatment suggestions have been incorporated into each chapter. The book will not, in general, instruct the reader in making choices among drugs or in devising pharmacological drug regimens. To

3

4 SAMMONS AND SCHMIDT

do so well requires a sound grasp of fundamental principles of pharmacology and psychopharmacology that cannot be imparted by this or any other single volume. Of course, clinical experience is the most basic pre- requisite to effective prescribing, and this can be acquired only by means of appropriately supervised direct experience. In the past, acquisition of such clinical experience was limited to psychiatrists and other medical practitioners. Now, however, a number of training programs have been initiated to train psychologists, advanced-practice nurses, and other nonmedical professionals in these skills-evidence that nonmedical professions are increasingly aware of the importance and value of education in psychopharmacology-

The book is organized by diagnosis. Psychologists wilI recognize that there are certain perils in this approach because of the limitations of syn- dromic categorizations of mental distress. Depressive disorders, for example, often have significant anxiety components, and psychologists have long been sensitive to the fact that patients and their difficulties cannot be reduced to Diagnostic and Statistical Manual of Mental Disorders- type (4th ed., D S M - N , American Psychiatric Association, 1994) checklists with rote treatment plans that are uniform for all. More than in any other health care field, the wisdom of the adage that to treat the patient, not the diagnosis, is apropos to mental health interventions.

Although this book is not a primer on psychopharmacology, each chapter provides a broad overview of current pharmacological interventions and often a preview of pending innovations in pharmacological treatment. For readers seeking an in-depth discussion of basic psychopharmacology or principles of psychotropic drug management, the following resources exist. Of the general clinical references designed to help the reader devise appropriate drug intervention strategies, those by Gelenberg and Bassuk (1997); Schatzberg and Nemeroff (1998); or Janicak, Davis, Preskorn, and Ayd (1997) are among the most complete. Readers interested in basic principles of psychopharmacology cannot do better than to add textbooks by Cooper, Bloom, and Roth (1996); Feldman, Meyer, and Quenzer (1997); or Bloom and Kupfer (1995) to their bookshelves. Stahl's (19961 book is a solid, uncomplicated general reference. Pagliaro and Pagliaro (1997,1999) also have added to the literature by providing textbooks of basic clinical psychopharmacology that are written from a psychological perspective.

Who Should Read This Book?

The primary audience for this book are practicing clinicians who seek to incorporate scientifically informed opinion into treatment planning and case management. Psychologists, counselors, and other nonmedical practitioners engaged in behavioral treatment who seek to understand more about the pharmacology and the combined treatment of specific disorders will find this book helpful. The book will be equally helpful to medical practitioners who seek to understand more about both combined treatments and behavioral or psychotherapeutic modalities, as well as those

INTRODUCTION 5

who wish to update their knowledge regarding current pharmacological treatments. Academic psychologists and their students may also find this book of interest, for many of the chapters are written by renowned experts in their fields and represent not only state-of-the-art reviews but also a keen vision of future research and treatment.

An emerging audience for this book is the small but growing cohort of psychologists who have completed specialized training in psychopharmacology. Such psychologists are currently rare, but numerous programs around the United States are now training psychologists to prescribe. The chapters in this volume will be of use to instructors and students in such programs in that they provide a truly psychological perspective on the prescription of psychotropics. By doing so, i t is hoped that this book will assist in the development of an academic model that, while providing psy- chopharmacological training of the highest caliber, is firmly grounded in the discipline of psychology.

Plan of the Book

Chapter 1, by Morgan T. Sammons, outlines some hypotheses as to why combined treatments have historically been neglected and offers some general clinical considerations for combining treatments. These general clinical guidelines are then expanded on in subsequent chapters that deal with specific disorders.

Ethical and professional issues involved in psychologists’ acquisition of prescriptive authority are addressed in the chapter 2, by Patrick H. DeLeon, Sharon E. Robinson Kurpius, and John L. Sexton. This contribution speaks directly to the experience of psychologists in their pursuit of prescriptive authority. Although members of other professions may not at first find the material contained in this chapter to be of direct applicability, closer inspection is warranted. The ethical principles outlined in this chapter are rooted in ethical principles for psychologists, yet they are universal in their application and are just as fundamental to good psychiatric or nursing practice. Members of nonmedical professions who seek to expand their authority to use medication also will profit from examining this chapter. DeLeon et al. discuss at length the findings of the recent Pew reports on the changing scope of practice of nonmedical professions. This provides a glimpse of the future landscape of health care and the nature of expanded service provision by psychologists, nurses, and other professionals whose practices have been constrained by tradition, but not by logic, from the provision of pharmacological services.

Chapter 3, by Martin M. Antony and Richard P. Swinson, and chapter 4, by Norman B. Schmidt, Margaret Koselka, and Kelly Woolaway-Bickel, are devoted to an exploration of anxiety disorders. As noted above, some controversy exists regarding the utility of pharmacological interventions in treating anxiety disorders because of the observed durability of behavioral techniques. Certain medications, however-notably, benzodiazepines, tricyclic antidepressants, and the selective serotonin reuptake in-


hibitors-have also proven to be powerful tools in treating numerous anxiety-based conditions. In the last several years a number of selective serotonin reuptake inhibitors and other newer antidepressants have received a U.S. Food and Drug Administration indication for obsessive- compulsive disorder (OCD), social anxiety and social phobia, and panic disorder, making them an increasingly viable treatment option for individuals who are not responsive to behavioral intervention. Antony and Swinson, in their contribution on OCD, demonstrate that both pharmacological and behavioral approaches are of significant value in a disorder that may be mediated by perturbations in serotonergic neurotransmission. They provide valuable outlines of both pharmacological and behavioral treatments that will be of interest to clinicians working with patients with OCD. They suggest caution in applying combined treatments, largely because the few studies that have been carried out to date do not demonstrate a clear-cut advantage for such treatments. Combined treatments may be of benefit as augmentation strategies, however, or when comorbid depression or other conditions complicate the clinical picture.

In chapter 4, Schmidt et al. cautiously explore the use of combined treatments in phobic anxiety disorders. Although they note that the majority of patients with such disorders have received both medication and behavior therapy, systematic study of these treatments together has been limited. In those studies that exist, a wide range of outcome is often reported. While pharmacological interventions have demonstrated efficacy, relapse is common on discontinuation. The authors note that the timing of interventions may be an important variable in treatment, as cognitive- behaviorally based strategies may be of assistance when using fading pro- cedures for drug treatment. They raise the notion of treatment specificity, that is, that subsets of symptoms of phobic anxiety disorders may be differentially responsive to either drug or nondrug treatment. This hypothesis requires further validation, but it seems likely that in syndromes such as bipolar disorder or schizophrenia differentially responsive symptoms exist. There is no reason to assume that symptom clusters or variable susceptibilities to a particular form of treatment do not exist for panic disorder and other phobic anxiety disorders.

Charles M. Morin, in his discussion in chapter 5 of insomnia and other sleep disorders, echoes a refrain that should be familiar at this point: that few evidence-based guidelines exist to aid the clinician in devising combined treatment strategies for this spectrum of disorders. No single approach is effective for all subtypes of sleep disorder. Pharmacological approaches are highly effective in the short-term treatment of insomnia, but tolerance to their effects, risks of dependence, and rebound on discontinuation mitigate against their prolonged use. Behavioral treatments, for individuals who respond to them, appear to be more robust. Morin suggests that combined treatments may be more effective if used concurrently -that is, an initial course of medication, coupled with behavioral management principles-but again cautions that the literature as yet provides scant sttpport.

Depression is the most commonly treated problem in mental health

INTRODUCTION 7

offices and among the most common presenting complaints in primary care clinics. Despite its commonality and the intensity with which it has been studied, treatment is unstandardized, and the ideological divisions among various forms of intervention are wide. This is no doubt partially caused by cultural conceptions of depression (see Healy, 1997) as well as by difficulties in capturing the experience of depression under the prevailing DSM-N-based nosological system. This clearly has limited the investi- gation of combined treatments, of which there are astonishingly few for such a common disorder, as the review in chapter 6 , by Jeremy W. Pettit, Zachary R. Voelz, and Thomas E. Joiner, Jr., attests. These authors nevertheless report that combined treatment studies carried out to date suggest a modest effect for this approach. Pettit et al. also note that most studies in this area have been carried out using medications that are no longer the initial treatment of choice (e.g., the tricyclic antidepressants). They note that depression is a multifaceted problem. Some patients may do well with unimodal approaches; however, evidence suggests that many may do best with combined treatments.

Current treatments for schizophrenia and other psychotic disorders reflect two dramatic changes in the 1990s. The first was the escalating de- emphasis on inpatient treatment in favor of shorter hospital stays and greater reliance on outpatient rehabilitation brought about by the advent of managed care. The second is the introduction of the atypical antipsychotic agents. As William D. Spaulding, Dale L. Johnson, and Robert D. Coursey note in chapter 7, on combined treatments in schizophrenia, these new drugs have enhanced the role of psychosocial rehabilitative efforts, because they hold the promise for long-term recovery rather than the symptom palliation afforded by earlier generations of antipsychotics. Whether the atypicals will fulfill this promise is as yet unknown, but it is clear that the contribution of psychosocial treatment in schizophrenia must be re-evaluated as necessary components in any treatment plan. As Spaulding et al. point out, the traditional psychiatric focus on symptom suppression in increasingly obsolete. A range of specific psychological and psychosocial interventions are available to assist people with schizophrenia in sustaining higher levels of recovered function, and it is more and more apparent that these interventions form a vital component of any comprehensive treatment plan. Spaulding et al. make a valuable contribution in terms of a treatment algorithm that may assist decision making in applying these interventions.

Perhaps the most strikingly successful example of combined treatments this book can offer is represented by chapter 8, Marc E. Mooney and Dorothy K. Hatsukami’s contribution on treatments for tobacco cessation. Nicotine dependence is a problem with strong biological and psychological correlates. Mooney and Hatsukami successfully demonstrate that interventions addressing both behavioral and physiological components are more likely to succeed than approaches that address only one facet of the disorder. Because the long-term consequences of nicotine dependence are severe, and because combined treatments are of demonstra-


ble robustness, psychologists should see this chapter as a true invitation to become more involved in the treatment of nicotine dependence.

Kathleen M. Carroll’s chapter 9 on treatment of substance abuse (other than nicotine dependence) demonstrates a more adjunctive, yet still important role for pharmacological intervention in the treatment of addic- tive behavior disorders. Although the implicit thesis in Carroll’s work is that behavioral principles are fundamental to successful management of substance dependence, the chapter also acknowledges the reality that no “magic bullet” exists. Clearly, the direction of the field is in seeking those appropriate combinations of treatment that best suit the individual suf- fering from substance abuse disorders (Boucher, Kiresuk, & Trachtenberg, 1998), and Carroll’s chapter is a useful guideline for clinicians attempting to put this philosophy into practice as well as a masterful review of the substance abuse treatment literature.

Obesity is another disorder with strong physiological and psychological substrates. Both pharmacological and behavioral treatments are still evolving for this pernicious problem, which is often accompanied by significant medical comorbidity. Recent well-publicized negative outcomes associated with the “phen-fen” regimen have given pause to advocates of pharmacological interventions and, although Carlos M. Grilo notes in chapter 10 that more recent innovations, such as the lipase inhibitors, are free of these negative side effects, other problems mandate that these medications be carefully deployed. Because of the risks associated with pharmacological intervention, and because the debate as to whether purely behavioral treatments provide equivalent outcomes to drug regimens is not yet settled, a conservative approach to pharmacological management of obesity seems prudent. Nevertheless, a strong case for combined treatments can be made in cases in which high body mass indexes or medical comorbidity exist. Because of the high rate of relapse that generally follows discontinuation of pharmacological treatment, it may be reasonable to argue that the choice rests not between medication or behavioral treatment but between combined treatments versus behavioral treatment alone.

James M. Meredith, Michael J. Lambert, and John F. Drozd present in chapter 11 a n outcomes assessment package that they have developed for use in a clinical setting with a focus on the Outcomes Questionnaire (OQ45.2; Lambert et al., 1996). Clinicians seeking accessible and clinically useful tools that have sound psychometric properties will find this chapter of particular interest. These authors have taken care in assembling a package of outcomes measures that largely conforms with the recommendations of the National Institute of Mental Health panel on clinical outcomes (Newman & Ciarlo, 1994). Readers will note that the measures are designed to be independent of treatment provided. This helps in meeting the requirement of clinical utility but makes it difficult to ascertain the con- tributions of drug and nondrug components of treatment. Are treatment- specific outcome measures necessary in clinical practice? This is a debat- able question. On the one hand, it can be argued that if one uses a combination of previously validated treatments that share as mutual goals

INTRODUCTION 9

the reduction of the same set of symptoms, then treatment-specific outcomes add little to good clinical assessment. On the other hand, this answer is not likely to satisfy those who seek to isolate those factors, or combinations of factors, that contribute most to good clinical outcome. This is an area that cries out sharply for further careful research.

Ideological divisions, poorly fitting research strategies, and a gap between science-based and ordinary clinical practice have impeded a more complete understanding of the mechanisms and effectiveness of combined treatments. At the most fundamental level, the question of whether they are more efficacious than unimodal treatments has yet to be definitively answered. Yet evidence in favor is slowly accreting, at least for certain disorders and, as the chapters in this book attest, progress in other areas, however slowly, is being made. Thorny practical and ethical problems remain: From a practical perspective, is it reasonable to hope that uniform clinical decision-making strategies for selecting combined treatments can be developed? Ethically, can such strategies be developed so that they do not repeat the mistakes of the past-primarily, an excessive reliance on psychotropic agents?

Newer research models specifically designed to address combined treatments will help answer these questions. In order to have an influence on practice, however, educators and trainers in psychology must adopt and disseminate new statistical and heuristic models for understanding combined treatments. If we do not train future psychologists, both academics and clinicians, to appreciate the value of a more catholic approach toward the treatment of mental disorders, we will needlessly constrain the ability of the field to advance and to offer the widest possible range of treatment options to those whom we seek to serve.

References

American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author.

Bloom, F. E., & Kupfer, D. J. (Eds.). (1995). Psychopharmacology: The fourth generation of progress. New York: Raven.

Boucher, T. A., Kiresuk, T. I., & Trachtenberg, A. I. (1998). Alternative therapies. In A. W. Graham, T. K. Schulz, & B. B. Wilford (Eds.), Principles of addiction medicine (2nd ed., pp. 371-394). Chevy Chase, MD: American Society of Addiction Medicine.

Cooper, J. R., Bloom, F. E., & Roth, R. H. (1996). The biochemical basis ofneuropharma- cology (7th ed.). New York: Oxford University Press.

Feldman, T. S., Meyer, J. S., & Quenzer, L. F. (1997). Principles of neuropsychopharmacol- ogy. Sunderland, MA: Sinauer Associates.

Gelenberg, A. J., & Bassuk, E. L. (Eds.). (1997). The practitioner’s guide to psychoactiue drugs (4th ed.). New York: Plenum.

Healy, D. (1997). The antidepressant era. Cambridge, MA: Harvard University Press. Janicak, P. G., Davis, J. M., Preskorn, S. H., & Ayd, F. J. (1997). Principles and practice o f

psychopharmacotherapy (2nd ed.). Baltimore: Williams & Wilkins. Lambert, M. J., Hansen, N. B., Umphress, V., Lunnen, K., Okiishi, J., Burlingame, G. M.,

& Reisinger, C. W. (1996). Administration and scoring manual for the OQ45.2. Steven- son, MD: Professional Credentialing Services.


Newman, F. L., & Ciarlo, J. A. (1994). Criteria for selecting psychological instruments for treatment outcome assessments. In M. E. Maruish (Ed.), The use ofpsychological testing for treatment planning and outcome assessment (pp. 98-110). Hillsdale, NJ: Erlbaum.

Pagliaro, L. A., & Pagliaro, A. M. (Eds.). (1997). The pharmacologic basis of psychothera- peutics: An introduction for psychologists. New York: Brunnerhlazel.

Pagliaro, L. A,, & Pagliaro, A. M. ( 1999). The psychologists’neuropsychotropic drug reference. New York: Brunner-Routledge.

Schatzberg, A. F., & Nemeroff, C. B. (Eds.). (1998). Textbook of psychopharmacology (2nd ed.). Washington, DC: American Psychiatric Press.

Stahl, S. (1996). Essential pharmacology. New York: Cambridge University Press.

1

Combined Treatments for Mental Disorders: Clinical Dilemmas

Morgan ll Samrnons

Believe those who are seeking the truth. Doubt those who find it. -Andre Gide

The absence of a compelling body of evidence on combined pharmacological and nonpharmacological treatments for mental disorders is perhaps the most striking feature of the mental health clinical research literature. This lack of data-particularly in an age of evidence-based practice-about what is arguably the most common form of treatment for mental distress suggests much about the degree to which guild and financial interests shape the pursuit of scientific knowledge. My first task in this chapter is to document the prevalence of combined treatments. I then examine the academic and political phenomena that have contributed to the paucity of data on combined interventions. Obstacles, surmountable or otherwise, to our understanding of these treatments are discussed (along with some occasional successes). I then turn to more practical matters, notably, how one might proceed in developing appropriate standardized protocols that clinicians can use when formulating and applying combined interventions. Because the literature is largely silent, it is difficult to formulate clear, systematic guidelines directing clinicians toward optimum combined treatment strategies. Some tentative guidelines are be offered, but it is ac- knowledged that the current state of understanding renders these guidelines aspirational and, it is hoped, ephemeral, in that directives that are more solidly grounded in science will be forthcoming.

A Failure of Investigative Models: Some Flaws, Fallacies, and Conundrums

Combined drug and nondrug treatments for mental distress are poorly represented in the research and clinical literature. Nevertheless, they are

The opinions expressed by this author represent his views as a private citizen and should not be construed as representing the official opinions or positions of the U.S. Navy or Department of Defense.

11

12 MORGAN T. SAMMONS

widespread in clinical practice, to the extent that they may be said to constitute the norm. A significant percentage, possibly even the majority, of all patients receiving services from a psychologist or other nonprescrib- ing mental health practitioner are also simultaneously receiving psychotropic medications, as demonstrated by a number of surveys of mental health service providers (“Mental health,” 1995; Sammons, Gorny, Zinner, & Allen, 2000; Chiles, Carlin, Benjamin, & Beitman, 1991). A further tell- ing indicator of the common nature of combined treatments is the frequency with which primary care practitioners, who are most likely to initially encounter and diagnose mental disorders, use both drugs and referral to mental health specialties. A recent survey demonstrated that 72.5% of depressed patients were given antidepressants, and 38% of these were also referred to a mental health specialist (usually a psychologist or social worker; Williams et al., 1999).

On the other hand, pharmacological treatment has become the mainstay of psychiatric service provision. Reporting on the National Ambula- tory Medical Care Survey data from 1985 and 1993-1994, Olfson et al. (1998) reported that at least one antidepressant was prescribed in 48.6% of all visits to psychiatrists in 1993-1994. Using the same data set, Pincus et al. (1998) discovered that, in 1993-1994, a visit to a psychiatrist specifically for depression resulted in the prescription of a psychotropic agent in 70.9% of cases. Because not all visits to psychiatrists are for depression, the total proportion of visits in which drugs were prescribed was undoubt- edly much higher. This assumption was confirmed by a survey of the practice of 148 psychiatrists in routine outpatient practice (West, Zarin, & Pincus, 1997). In this survey, 90% of all patients of psychiatrists were prescribed at least one psychotropic medication (the mean number of medications per patient was 1.8). In a further analysis of this data set, Pincus et al. (1999) reconfirmed that, in 1997, approximately 90% of patients of psychiatrists surveyed were taking medications. As the authors noted, this was a sizable increase since 1989, when 54.5% of psychiatric patients were prescribed medication. Pincus et al. (1999) also found that 55.4% of out- patients reported on in this survey received both medication and psychotherapy, with psychotherapy being provided either by the psychiatrist or another professional. It is apparent, then, that pharmacotherapy is the mainstay of current psychiatric practice but, even so, the majority of patients also receive psychotherapeutic services. Zit0 and colleagues (2000) also documented a n extraordinary rise in the rate of prescriptions of psychotropics to preschoolers during the 199Os, indicating that the overpre- scription phenomenon is hardly limited to adult populations.

Unfortunately, the pervasiveness of combined treatment is poorly documented in clinical research, and its mechanisms and effectiveness remain the focus of controversy. This in large part may be because of the power of the controlled clinical trial as an investigatory heuristic. Although the benefits of controlled clinical trials cannot be disputed, in certain respects this model has led to an investigative approach that does not capture well the nuances involved in combined treatment. The literature is replete with reports of single-modality, placebo-controlled outcome studies, such as the

CLINICAL DILEMMAS 13

effectiveness of cognitive-behavioral models in treating depression. Also, a reasonable number of comparative-treatment outcome studies exist for most major mental disorders. These “horse race” studies often involve head-to-head comparisons of unimodal pharmacological and psychological interventions. Although they have become somewhat less common in recent years (Beitman, 1991), these studies continue to be highly represented in the literature. At the same time, trials of combined treatments are scarce. Only a handful, of variable quality, exist for most disorders.

In part, this situation has been perpetuated by professional biases. Psychologists and nonmedical researchers may have a vested interest in demonstrating the superiority of nonpharmacological techniques. On the opposite side, psychiatric researchers, particularly those with a biological orientation, may tend to champion pharmacotherapeutic strategies. These dichotomous conceptualizations of interventions lead to difficulties in research design and provide a source of investigatory bias that can consid- erably influence outcome. Sources of investigatory bias are difficult to isolate precisely but are reflected by practices such as comparing the treatment being studied against one that appears equivalent but in reality is unequal. One common example of this in drug studies is the strategy of comparing a new drug against an older agent that is effective but has a less favorable side-effect profile. This practice has been found to be ex- tremely common in schizophrenia research (Thornley & Adams, 1998). Re- searchers’ preference for, or allegiance to, one form of treatment over another may also lead to the less favored treatment being inadequately implemented during a clinical trial (Jacobson & Hollon, 1996). A further difficulty in research design is not directly related to hidden researcher bias but is endemic in much of mental health research today. This is the familiar difficulty encountered when efficacy, rather than effectiveness, studies are performed. Efficacy studies, which I discuss in more detail later, comprise the bulk of the scientific knowledge base in mental health research. These studies, usually based on comparisons of two reasonably pure treatments applied in sterile research environments to participants who resemble each other as much as possible, result in outcomes that are poorly generalizable to the everyday treatment setting. As compared to effectiveness studies (examinations of how patients respond to treatments applied in the field; Seligman, 19951, efficacy studies have limited ability to satisfactorily inform clinicians or patients as to optimum choices among treatments (Roland & Torgerson, 1998).

Controlling Bias in Research and Practice

In light of findings that neither psychotherapeutic nor pharmacotherapeutic approaches are superior in the treatment of at least the most common form of mental distress, preference emerges as a key, if not decisive, factor in determining selection of treatment. Preference may be expressed by either the clinician or the patient. Patient choice is an important variable in determining positive outcome, but patient preferences are probably in-


fluenced by clinicians to a greater extent than is realized. Strongly held opinions about what is best for patients not only prevents clinicians from uncritically examining the data and values that shape their assumptions, but they also may make clinicians insensitive to the rights of patients to disagree (Woolf & Lawrence, 1997). In the field of mental health, clinicians are peculiarly positioned to interpret differences of opinion between therapist and patients as being rooted in psychopathology (i.e., resistance), rather than as an issue of patient choice:

Some patients want only medications and others want only psychotherapy. Those who ask for medications only may simply want immediate relief and not care what the means is. On the other hand, those who want psychotherapy only may reject medication out offear of some external control, preferring instead a sense of personal control. Al- though each of these positions may be considered resistance to oppo- sitely oriented psychiatrists, they are more specifically resistances to the bias of the psychiatrist. (Beitman, 1991, p. 26)

The obvious challenge is to create a system in which data, and not bias, drive treatment recommendations. With such evidence, the clinician will be able to offer the patient expert advice as to the form of intervention best suited for the presenting complaint. By careful, unbiased education, patient attitudes may be changed so that they can be steered toward whatever form of intervention has been demonstrated to be most effective. The right of a patient with disabling symptoms of anxiety to demand immediate relief in the form of an anxiolytic medication must be respected, not challenged. At the same time, the clinician must take pains to educate the patient that this relief is likely to be short term and evanescent once the medication is discontinued. Such patients should be given impartial information as to the availability of potentially more effective treatments leading t o longer lasting relief. This information should include a discussion of whether nonpharmacological treatment can be used in combination with medication; as a substitute for it; or if the medication will interfere with the process of behavioral treatment, as may be the case when benzodiazepines are used in conjunction with exposure-based treatments for phobic anxiety.

It is obvious that we are far from reaching the ideal of providing patients with unbiased, purely objective informed consent. Practically, this state of reason is probably impossible to attain. Biases, expectations, and differences in information processing continually affect interchanges between therapists and patients (Redelmeier, Rozin, & Kahneman, 1993). The goal should not be to eliminate such biases but to minimize their influence by making them explicit to both patient and therapist, so that each may judge the effects of their beliefs on choice of treatment.

The Burden of Reductionistic Thinking

Subtle investigator bias resulting from dichotomous thinking about mental health interventions is but one complicating factor that has led to com-


bined treatments being understudied. Another factor that has significantly influenced research patterns has been the quest to identify, with increasing specificity, “cures” for mental disorders. This search represents some- thing of a conundrum, which can be outlined in broad strokes as follows: Psychological distress is a heterogeneous and nonspecific concept, and its experience is unique to each sufferer. One can define, albeit in rather neb- ulous terms, some of the features that separate one form of psychological distress from another, but it remains true that most people with schizophrenia, or most depressed patients, share in common only the most obvious features of their diagnoses. Nevertheless, the aim of much of mental health research in the past 50 years has been to search for increasingly specific remedies. We are therefore placed in the awkward position of pos- iting molecular cures for molar concepts that are heterogeneous, nonspecific, and experienced in an absolutely unique manner by each sufferer.

The past 50 years of mental health research has led to the successful development of many specific pharmacological and psychological treatments that have improved patient outcomes (Michels, 1999). At least in the short term, specific pharmacological interventions do assist many patients in coping with the more disabling aspects of their illness, sometimes dramatically so. Yet there is also evidence that these increasingly specific results do not translate into lasting improvement. Rates of successful treatment for schizophrenia have not appreciably changed in the past 100 years (Hegarty, Baldessarini, Tohen, Waternaux, & Oepen, 19941, despite the synthesis of effective antipsychotic drugs. New-generation antidepressants, such as the serotonin reuptake inhibitors, have not resulted in improved long-term remission rates, neither have increasingly specific psychological treatments. In the well-known (if not overstudied) Treatment of Depression Collaborative Research Project (Elkin et al., 1989), recovery rates at 18-month follow-up did not differ among any treatment. Recovery ranged from 19% for clinical management plus imipramine to 30% for cognitive-behavior therapy (CBT; Jacobson & Hollon, 1996), a less-than- splendid showing for any treatment. To a large extent, then, specificity and success do not correlate well.

Paradoxically enough, increasing the specificity of treatment has constrained our ability to perform certain types of research. Because one can demonstrate the success of specific treatments in short-term (although rarely in long-term) outcome studies, we have greater difficulty justifying the application of combined treatments. Essentially, the issue is the ability to justify a more complex, possibly more expensive treatment when sim- pler and cheaper remedies have been shown to be of utility. Is it ethical to impose unproven, costlier combinations on patients when less complicated alternatives, already shown to be of value, exist? This question is subject to considerable debate and arises in numerous examples throughout this chapter.

The issue of specificity pertains to diagnoses as well as treatment. I t is a grave error to assume that, once having made a Diagnostic and Sta- tistical Manual of Mental Disorders-type (DSM) diagnosis, the treatment becomes uniform. Hohagen et al. (1998) demonstrated, for example, that


patients with DSM-III-R (American Psychiatric Association, 1987) obsessive-compulsive disorder (OCD) did best with unimodal therapy (behavioral treatment) if their symptoms were primarily compulsive but did best with combined medication and behavior treatment if their symptoms were primarily obsessive. Along similar lines, Wells and Sturm (1996) found that addition of minor tranquilizers to antidepressant therapy did nothing to improve outcomes in the treatment of major depressive disorder. Yet it is clear that a subset of patients with major depression present with significant anxiety symptoms. When these symptoms are appropriately managed with a short-term course of benzodiazepines, outcome is improved (Buysse et al., 1997; Smith, Londborg, Glaudin, & Painter, 1998).

This introduction should remind the reader that in spite of the high prevalence of combined treatment in clinical practice our knowledge of combined treatments is poor. They may not work as well as single- modality treatments for some disorders; they may provide more rapid or lasting relief in others. Because combined treatments are often not sup- ported by the current literature, clinicians should be circumspect in devising such treatments for their patients. At the same time, clinicians should be careful to balance the needs of individual patients against the results of large-scale studies or meta-analyses, for these are poor predic- tors of individual response in the clinical setting (cf. Klein, 1998). For most conditions, single-modality treatments should be attempted before combined treatments are implemented and, for all conditions for which it has found to be effective, psychosocial treatment should be included in the treatment plan.

Unresolved Issues in Combined Treatments

There is bountiful support that psychopharmacotherapy provides generally incomplete and temporary relief from mental distress. There is also equally convincing evidence that credible forms of active psychotherapy are generally indistinguishable in terms of efficacy. Long-term outcome data pay no compliments to either approach. Thus, advocates of neither biological nor psychosocial approaches have much in the way of substan- tive data to support claims that theirs is the preferred method of intervention. Conflicts between various schools of mental health practitioners are, then, generally based in ideology (Merman, 1991) rather than science. Because ideological allegiances have limited the study of combined treatments, clinicians lack data to guide their application. Some of the more important factors that remain poorly understood are the timing of particular components of combined treatments, our understanding of the nonspecific factors associated with any component of treatment, and how decisions about drug or nondrug treatment can be better standardized. It is to these issues that I now turn.


Timing of Interventions

If combinations of drugs and verbal therapy are used, when is it reasonable to introduce each component into the treatment plan? This largely unexplored area is of importance in determining when and if a combined strategy is indicated and how combined treatments are optimally applied in clinical settings. Miller and Keitner (1996) provided a thoughtful review on the subject and suggested that at least three strategies are possible. The first involves administering all treatments simultaneously. Providing all treatments concurrently would ensure that the patient has been ex- posed to all elements of potential value. This approach, however, is both costly, because greater resources are expended, and inefficient, because it is impossible (at least given the current state of understanding) to identify a priori those patients who will respond to a specific component of treatment.

A second alternative is the sequential model, wherein additional treatments are proffered on the basis of response or lack of response to previous interventions. Miller and Keitner (1996) noted that this technique is already almost universally used in drug treatment-doses are increased, or different drugs are attempted, if the first medication has proven ineffec- tive. This, as the authors noted, is a more parsimonious and potentially cost-effective approach in that additional interventions are offered only if previous ones have failed. A potential drawback to this approach is that any beneficial synergistic effects of offering treatments together might be either deferred or lost. In addition, dose-response relations evidently exist for psychotherapies (Howard, Kopta, Krause, & Orlinsky, 1986) as well as pharmacotherapies, and this effect could be lost by adding psychotherapy later in the treatment course (i.e., too little, too late), or it could be ob- scured by the addition of a drug treatment.

Third, Miller and Keitner (1996) proposed a “matching” strategy, wherein various single or combined treatments are offered on the basis of an assessment of the patient’s identified deficits or resources. This, they noted, is also a cost-effective model, but if treatments and patients are matched incorrectly, outcomes will be suboptimal. Because, as observed previously, one cannot easily determine in advance those components of treatment to which individual patients are likely to respond well, this may be the least preferred of the strategies for combining. Using depression as an example, it is often very difficult to clinically determine when presenting symptoms represent acute onset of a major depressive episode, an adjustment disorder, or an acute stress reaction. Although history may be of some assistance in distinguishing among disorders that may require longer term pharmacological management and those that are expected to resolve with brief treatment, this is not always the case. Suicidal ideation as a presenting complaint may result from cognitive factors (hopelessness); alternatively, patients may consider suicide as an escape from intolerable neurovegetative signs, such as severe insomnia or autonomic arousal. The dilemma here is whether to initiate a course of antidepressant therapy immediately or to see if the patient’s symptoms will respond to several


closely spaced sessions of psychotherapy or environmental manipulation. Delaying antidepressant treatment may be deleterious, given that patients will in any case experience a 3- to 6-week time lag in onset of antidepressant effect. Initiating treatment immediately, however, may commit the patient to an unnecessary course of medication. It is perhaps best to tem- porize in these situations. Some experts have recommended that, in the case of milder, less chronic, nonpsychotic depression, an extended evaluation of two to three visits be undertaken to determine those patients who will remit with nonspecific treatment alone (Depression Guideline Panel, 1993). If a patient does not respond to closely spaced therapy sessions (perhaps augmented with short-term use of a benzodiazepine to address symptoms of insomnia and autonomic arousal; Smith et al., 1998), then delay in initiating a course of antidepressants is not likely to be of lasting harm.

In many combined-treatment outcome studies, both treatments have been initiated simultaneously at the beginning of treatment (Rush & Hol- lon, 1991). Rush and Hollon (1991) suggested that either could be added at any point in treatment without altering the modality already used. This statement may be true in the context of augmenting suboptimal responses to unimodal treatments (a reasonably well-studied maneuver). For instance, it is commonly recommended to add psychotherapy to a medication regimen if an inadequate response is present after 6-8 weeks of treatment. By using this strategy7 the additive effect of combined treatments can be estimated, but no knowledge is gained about the synergistic effects of two separate modalities applied simultaneously at some point in the treatment course, or whether reversing the order of the treatments applied would be more effective. Because no clinical outcome data exist to guide clinicians on this point, it is suggested that the following questions be asked when considering the timing of combined treatments.

First, has an adequate period of observation and assessment been ac- complished? Patients presenting in acute distress present diagnostic dilemmas. A moderate to severe adjustment disorder with depressed mood may be indistinguishable from a n acute stress disorder or the acute onset of a major depressive episode. Patients may demonstrate a rapid response to psychotherapy or environmental manipulation for the first two conditions and may not require initiation of pharmacotherapy. The risks of delaying treatment in a medication-responsive condition must be carefully weighed against any risk involved in the administration of drugs.

Second, have unimodal treatments already been considered or implemented? In general, pharmacotherapy alone is less effective than psychotherapy alone, especially in cases of treatment-resistant or chronic depression or when Axis I1 pathology or other conditions complicate the clinical picture.

Third, do contraindications exist to the use of combined modalities? Examples would be the use of a benzodiazepine during exposure-based therapy for phobias (Barlow & Lehman, 1996) or the use of relatively toxic agents, such as the tricyclic antidepressants or lithium in borderline patients or others with chronic suicidal or parasuicidal behaviors (Dimeff,


McDavid, & Linehan, 1999). There also may be medical contraindications to the use of pharmacological treatments, such as histories of cardiac difficulties in patients taking antidepressants. Although few psychotropics have been definitively linked to fetal abnormalities (Koren, Pastuszak, & Ito, 1998), research in humans is perforce limited. Some experts have recommended that women who are pregnant or contemplating pregnancy stop using antidepressants and anxiolytics unless a threat to the mother, such as suicide, exists (Diket & Nolan, 1997). This opinion is not held by all experts. Kulin et al. (1998) found no increased risk of major congenital malformations associated with antidepressant use in pregnancy in a pro- spective, controlled trial. Treatment of psychological disorders in the postpartum period also is understudied. The most common psychological problem in the postpartum period is depression, but a recent review identified only one controlled trial of antidepressants (Cooper & Murray, 1998). In the trial in question, both fluoxetine and counseling were found to be effective in treating postpartum depression (Appleby, Warner, Whitton, & Faragher, 1997). Numerous psychotropics are excreted in breast milk, but their effects on neonatal development are unknown (Stowe, Strader, & Nemeroff, 1998).

Fourth, for some conditions, in some individuals, combined treatments may represent optimum therapy, such as in bipolar disorder (Sachs, 1996); some forms of depression (Thase et al., 1997); for smoking cessation (Hat- sukami & Mooney, 1999); and, in all probability, psychotic disorders, such as schizophrenia (Rosenheck et al., 1998; Spaulding, Johnson, & Coursey, chapter 7, this volume). Does the patient manifest characteristics that have been demonstrated to be amenable to combined treatment? It is important to understand that these characteristics are fluid, will vary throughout an episode of illness, and must be reassessed on a ongoing basis. Significant depression, for example, may be complicated by numerous manifestations of anxiety early in the treatment course. Because of the delay in onset of antidepressant drugs it is important to recognize and treat these symptoms (Smith et al., 1998).

Fifth, has the patient’s history of response to either psychotherapy or pharmacotherapy been elicited? Patients whose initial response to pharmacotherapy has been positive may still require the addition of psychotherapeutic components. There is some evidence that exposure-based treatments can assist patients who initially used benzodiazepines to ob- tain relief from panic disorder. Benzodiazepines are effective in controlling the acute symptoms of panic but tend to provide long-term relief only with continued use. Risks of dependence (although probably overstated; Shader & Greenblatt, 19931, and the propensity for anxiolytics to interfere with exposure-based training, have led to recommendations to limit their use in the treatment of panic disorder. Bruce, Spiegel, and Hegel (1999) found that when anxiolytic agents are used, patients treated with CBT were significantly more able to discontinue alprazolam and remain symptom free at 2- to 5-year follow-up than those treated with standard management. Thus, a combination of pharmacological approaches, to ameliorate acute symptoms of the disorder, and psychotherapy, to provide long-term


relief, may be an appropriate strategy in panic disorder, although further study is required before this can be recommended with certainty.

Finally, what treatment modality does the patient desire? Has he or she been given adequate informed consent about the relative efficacy of either or both treatments? Integrating pharmacotherapy with psychotherapy early in the treatment course ideally will sufficiently reduce the more florid symptoms of a mental disorder to the point that the patient is able to effectively engage in a psychotherapeutic relationship (Herman, 1991). If this course is agreed on, patients must understand not only the risks and benefits associated with both pharmacotherapy and psychotherapy but also that the ultimate goal may be to withdraw the pharmacological agent prior to termination of therapy.

The Elusive Algorithm

During the 199Os, a number of attempts have been made to formulate rational prescribing strategies for psychotropics. In response to an emphasis on evidence-based practice and a need to manage rising health care costs, clinical guidelines have become increasingly common. Clinical guidelines are ideally evidence based, but many remain based on expert consensus or opinion (Woolf, Grol, Hutchinson, Eccles, & Grimshaw, 1999) and thus may not represent truly science-informed practice. Also, the evidence that underlies clinical guideline recommendations is intentionally biased toward highly controlled, diagnostically selective, randomized clinical trials (Shekelle, Woolf, Eccles, 8z Grimshaw, 1999); these generally take place in tertiary-care facilities with research capabilities. Such results likely do not translate perfectly to general treatment settings (Hay- cox, Bagust, & Walley, 19991, and their applicability in such settings has been challenged (Rosser, 1999). For example, the American Psychiatric Association’s practice guideline for major depressive disorder (Karasu et al., 1993) has been criticized for, among other deficits, undervaluing the efficacy of cognitive therapy and overstating the value of combining behavioral or brief psychodynamic therapy with medication (Persons, Thase, & Crits-Christoph, 1996).

One common method to standardize treatment is the development of formal algorithms. These are evidence-based guidelines providing treatment options for clinicians through a n episode of care. In general, commonly used drugs at low doses are selected first, with suggestions for use of drugs from other classes or other interventions should the disorder prove resistant. Algorithms have been developed for the treatment of schizophrenia (Pearsall et al., 1998) and major depression in primary care (Trivedi et al., 1998). One problem encountered in the development of algorithms is that the strength of the underlying evidence is often not very great. This is especially the case when new agents for which little clinical experience has accrued (such as the novel antipsychotics) are incorporated into an algorithm. In such instances conclusions may depend heavily on short-term, industry-funded trials (Pearsall et al., 1998).


Another problem associated with algorithms is their lack of ecological validity. Although combined treatments are common in routine practice, few algorithms address combined treatments, because these are rarely the subject of randomized clinical trials in tertiary-care settings. For example, Trivedi et al. (1998), in devising their treatment algorithm for depression in primary care, avoided any mention of referral for psychotherapy. A primary-care physician using such an algorithm to treat depression would have no prompt as to when or if a patient should be referred for psychotherapy. This is a particularly distressing oversight given the amount of evidence that psychotherapy is at least of equal efficacy (and, more con- troversially, occasionally superior) to pharmacotherapy in the treatment of depression (Munoz, Hollon, McGrath, Rehm, & Vanden Bos, 1994; Mur- phy, Carney, Knesevich, Wetzel, & Whitworth, 1995; Rush & Hollon, 1991). A solution to the current lack of ecological validity in many evidence-based guidelines would be the development of practice research networks. Such networks would enhance the ability to perform clinical trials in the primary care setting (Nutting, Beasley, & Werner, 1999) and would provide a mechanism for the systematic collection of data from potentially large numbers of participants in environments closely resembling actual practice conditions, where combined treatments are more likely to be prescribed.

Expert-consensus guidelines differ from evidence-based guidelines in that, as their name implies, they rely on the opinions of recognized spe- cialists in the treatment of a particular disorder. Recommendations are therefore more likely to represent current standards of excellence in practice rather than treatments suggested by randomized trials. Recommen- dations of experts, however, may be even more subjective than evidence- based guidelines, and they are less likely to be multidisciplinary, an important element influencing the acceptability of recommendations (Shekelle et al., 1999). As noted above, past recommendations by expert panels of psychiatrists have downplayed the effectiveness of psychotherapeutic intervention. An encouraging recent development is the trend to include nonpharmacological treatments as first- or second-line interventions for various disorders, such as that for OCD (March, Frances, Car- penter, & Kahn, 1998).

Evidence-based guidelines are increasingly common, and they represent a laudable attempt to match clinical practice with the best of research knowledge. As the shortcomings already discussed suggest, however, guidelines are no panacea. Like other forms of research, they may not be appreciated or implemented by clinicians. Guidelines have also been criticized because they have failed to take into consideration the costs of treatment, although there is some evidence that this is changing (Dean, 1999). In the final analysis, algorithms or guidelines for treatment of mental disorders may fail because both the manifestations of most mental disorders and the major effects of treatment are so nonspecific as to defy quan- tification in the form of an algorithm or guideline.

This problem is exemplified by our ambiguous understanding of the biology of depression and the wide variety of treatments for it. No theory


advanced to date can adequately explain what, if any, biochemical pertur- bation leads to the subjective experience of depression (Valenstein, 1998). It should therefore be no surprise that much remains to be understood about the pharmacology of antidepressant agents or that any single ex- planation of their mechanism of action is satisfactory (Shader, Fogelman, & Greenblatt, 1998). Why so many agents with differing or even opposing mechanisms of action produce an antidepressant response remains a n un- answered question (Hollister & Claghorn, 1993).’ Also, response to all drugs that are antidepressants is more or less the same. A depressed patient is just as likely to respond to fluoxetine as to amitriptyline or nefazodone. Manufacturers of antidepressants often attempt to distinguish their product by their neuroreceptor selectivity-whether a drug is more active on serotonergic or norepinephrine-containing neurons, for example. Although these claims reflect true pharmacological differences between antidepressant drugs, clinically all will produce the same degree of improvement, at least insofar as group data are concerned. The selective serotonin reuptake inhibitors (SSRIs) and other new antidepressants have superior side-effect profiles over older drugs, but remission rates have not improved (Burke & Preskorn, 1995). This may be, as Burke and Preskorn (1995) speculated, because some forms of depression are not responsive to pharmacotherapy, or because the mechanisms of action of available drugs are not appropriate for all subtypes of the disorder. Regardless, “there are no convincing data to suggest that regulations of adrenergic or serotonergic receptors per se [are] responsible for the therapeutic effects of antidepressant drugs” (Hyman & Nestler, 1996, p. 160). In clinical terms, this problem is illustrated in the algorithm by Trivedi et al. (1998). For a case of uncomplicated nonpsychotic major depression, SSRIs, nefazodone, bupropion, venlafaxine, moclobemide, mirtazapine, and the tricyclics are all listed as potential first interventions-a range of options that is hardly likely to satisfy a practitioner looking for algorithmic guidance on optimum drug strategies.

The smorgasbord of pharmacological alternatives that exists for treatment of most mental disorders may be said to represent for psychopharmacotherapy what the “dodo-bird effect” represents to psychotherapy; that is, all credible therapies tend to result in significantly greater improvement than do sham or placebo therapies, and there is little to distinguish one credible therapy from another. The dodo-bird hypothesis has recently been reconfirmed for psychotherapy (Wampold et al., 1997) and, because all credible antidepressants tend to (a) result in greater improvement than do other medications used for the same purpose and (b) result in approx-

‘In spite of this it is incontestable that certain medications have a specific antidepressant effect that can be behaviorally measured and can persist over time. Chronic administration of a n antidepressant compound to a severely depressed individual will have salutary effects that are distinct from those of a placebo (although the placebo may also have beneficial effects of its own). Chronic administration of a benzodiazepine, however, to a similarly depressed individual is not likely to result in a significant degree of improvement (Wells & Sturm, 1996). Thus it is clear that antidepressants differ not only from placebo in their effects, but they also differ from other classes of drugs.


imately the same rates of improvement, a dodo-bird effect can also be posited for drug treatment. If one accepts the argument that the dodo-bird principle applies to pharmacotherapy, which is buttressed by Kirsch and Sapirstein’s (1998) finding that the vast majority of the effects of antidepressant drugs are nonspecific effects, then it makes little sense to develop algorithms intended to standardize their use. Nonspecific effects are difficult to incorporate into formulaic treatment strategies. This is not, however, to suggest that nonspecific effects are bereft of therapeutic benefit; far from it. One does not use specificity as a measure of how effective a n intervention is. We cannot isolate the specific active components of CBT or other psychotherapies (Ablon & Jones, 1999; Jacobson et al., 19961, but CBT is an effective psychotherapy nonetheless, and to abandon it would disserve many patients. Neither should we abandon pharmacotherapy because we cannot identify active components of antidepressant treatment. As previously argued, specific treatments in fact may not be very good for many mental disorders, because they are multidimensional and, for each sufferer, uniquely experienced.

In the end, the search for an effective, universally applicable algorithm for the treatment of common mental disorders may be doomed to failure, inasmuch as it seems unlikely that the specific effects of any kind of treatment (pharmacological or not) will soon, if ever, be elucidated. Given this situation, nomothetic algorithms make little sense. As Woolf and Lawrence (1997) observed,

universal recommendations only make sense when there is little doubt across preference groups and risk profiles about the trade-off between benefits and harms . . . when what is best for one individual also is clearly best for another. If, however, the relationship of benefits to harms is uncertain and the “best choice” is a matter of personal values, family history, and other risk factors, a single policy for everyone is improper. (p. 2106)

These cautions, although written in reference to breast cancer, are highly applicable to any attempt to create algorithms for mental distress, because personal values and family history are integral not only to the understanding of these disorders but also often to their genesis.

Proceeding in the Face of Uncertainty

It is evident that, outside of clinical trials, we have much to understand about the standards of delivery and effectiveness of both pharmacological and psychotherapeutic treatments. Nevertheless, there is some guidance, if only aspirational in nature, that can assist the clinician in selecting the appropriate form of unimodal or combined treatment. Thus, an attempt can be made to assimilate the data covered previously in this chapter into some tentative recommendations to guide the selection and timing of treatments. The reader looking for specific intervention strategies is likely


to be disappointed; the preceding discussion should have educated any such seekers that we are still far from being able to delineate specific treatment recommendations.

The first task, regardless of whether unimodal or combined interventions are used, is the establishment of a n accurate, objective, and comprehensive diagnosis. A firm diagnostic base is necessary before proceeding with any specific intervention. I t is necessary to be alert to the presence of comorbid symptoms and evidence of dysfunction in spheres not addressed by DSM-based diagnoses when making the initial assessment. Here the clinician must use patient-based measures as well as clinical measures to achieve a multifactorial baseline. Consider ecological, inter- personal, intrapsychic, developmental, and biological variables. A small number of medications-for instance, some antihypertensives, steroids, or antineoplastic agents (Charney, Berman, & Miller, 1998)-have been at least tenuously linked to depression, and a careful documentation of a patient’s past and current pharmacological regimen is mandatory. This should include an accurate tally of any over-the-counter, alternative, or herbal medications, or nonprescribed psychotropics (e.g., marijuana, LSD, stimulants, opiates) taken by the patient. Although somatic causes for depression are reasonably uncommon, the patient should be assessed for the presence of disease states, such as hypothyroidism, that have been associated with depression. Biological and physical workups may be especially productive in the cases of individuals with known risk factors or in cases of new-onset symptoms in older individuals with no prior history and no adequate psychosocial explanations for their current presentation. No blanket physical or laboratory workup can be recommended; these will depend on the patient’s clinical presentation.

The patient’s full involvement in the treatment plan must be solicited. Ethically, this is a necessity and, practically, there is evidence suggesting that patient involvement in treatment planning improves outcome (Rich- ards, 1998). No treatment, psychological or pharmacological, should be offered without the provision of all elements of informed consent. For psychological therapies informed consent should include a description of the type of therapy, evidence supporting its effectiveness (including its effectiveness compared to no treatment at all), the demands it will impose on the patient, expected outcomes, and the frequency and cost of therapy sessions. Where evidence for the effectiveness of drug treatment exists, this alternative should also be explained to the patient. For pharmacological therapies, informed consent will include a discussion of the risks and benefits of the medication, as well as psychological treatments that may be equally effective, either singly or in combination with the drug treatment. In cases for which limited or no empirical support exists for pharmacological treatment, such as in the case of antidepressants in the treatment of depressed children and adolescents, this information should be clearly explained and available alternatives offered. The general finding that pharmacotherapies are less effective when offered without psychological interventions should be explained to the patient, and appropriate psy- &ological, behavioral, or psychosocial modalities (not necessarily indud-


ing formal psychotherapy) should be added to the treatment plan. In instances in which there is no clear evidence that one form of treatment is superior to another, the clinician must be observant of his or her inherent biases and serve as a candid, neutral source of expert advice. Strength of the evidence supporting both treatments, and the difficulties involved in translating research findings to the clinical setting, should be explained to the patient, who should then be allowed to select his or her treatment of choice. In many instances, patient choice will determine what treatment is to be applied, and clinicians must scrupulously adhere to their respon- sibility to provide the most expert and impartial knowledge in assisting such choices.

Observation of the nonspecific effects of intervention may assist in treatment decision making, especially decisions regarding pharmacological treatment. If the disorder improves during the assessment phase, as may be the case in the presence of a supporting and empathic therapist, a headlong rush for the prescription pad is unwise. As has been noted, some guidelines suggest that in milder, nonpsychotic or nonrecurrent de- pressions an extended (two- to three-session) evaluation may identify patients who will respond to nonpharmacological interventions (Depression Guideline Panel, 1993). When antidepressants are used, an excessively rapid or complete response that is not in keeping with the well-established delays in onset of effectiveness suggests that nonspecific or placebo effects are operating. These data can guide the clinician’s subsequent pharmacological maneuvers and may indicate that pharmacotherapy might be safely discontinued at a relatively early point in treatment.

The effects on the therapeutic relationship of prescribing should also be considered, although admittedly there are essentially no solid data to guide clinicians here. In analyzing the limited literature on this topic, Klerman (1991) noted that most such effects were presumed to be negative. For example, the act of prescribing was thought to either diminish the effectiveness of psychotherapy by making the patient dependent on an authoritarian prescribing figure or to lead patients to forgo potentially more lasting change in favor of symptom reduction. These concerns, although worthy of some consideration, should not be dwelt on excessively. There are certain instances in which the addition of drugs may reduce the effectiveness of psychotherapy (the earlier-cited example of the use of an- ziolytics with exposure-based therapies). Otherwise, most opinions on the negative influence of drugs on psychotherapy are speculative.

This said, it is important not to dismiss the use of pharmacological agents when they are necessary. If levels of distress are high, particularly if they effectively preclude the establishment of a sound therapeutic liai- son, consider the immediate use of combined treatment with the appropriate pharmacological agent. In catatonic or psychotic depression, other psychotic states, mania, and cases of extreme anxiety or agitation, pharmacotherapy is likely to be the first-line intervention. Even in such cases, a strong therapeutic alliance should be maintained throughout the hos- pitalization, and these interventions should never be offered without ongoing psychosocial intervention. Pharmacological treatments should never


be offered as sole interventions. Schooler and Keith (1993) reviewed the literature on treatments for schizophrenia and found considerable evidence that currently available antipsychotic medications are effective in controlling both acute symptoms as well as in delaying or preventing relapse. Psychosocial treatments, including individual psychotherapy, individual social skills training, and family therapy, had additive or interactive roles with medication in treating both acute and long-term phases of the disorder. Family therapy was found to be a particularly effective intervention in long-term treatment. Janicak, Davis, Preskorn, and Ayd (1997) performed a meta-analysis of seven studies that examined combined psychosocial and pharmacological treatment in people with schizophrenia. Patients who received both psychoeducation and family therapy did significantly better than patients who received drugs alone.

In general, the literature suggests that unimodal treatments should be offered prior to combined treatments because of current limited evidence that combined treatments offer significant added benefit. In imple- menting this standard the reader should heed the caution, extensively discussed throughout this chapter, that combined treatments are more common than not and that current investigative models in all probability have underestimated the true effect of combined treatments. Where there is clear evidence that unimodal treatments may provide superior outcome, however, these should be implemented first. Such examples include many of the anxiety spectrum disorders. Behavioral treatment or CBT is the initial treatment of choice for OCD, panic disorder, and specific phobias. If adequately provided psychological therapy yields a n incomplete response, it should not be discontinued, but pharmacotherapy should then be added into the treatment plan. Benzodiazepines should be largely avoided, especially in exposure-based therapies, because of their inhibition of learning responses and the risks of pharmacologic dependence.

Both pharmacological and behavioral strategies must be continually monitored and titrated throughout the treatment course. I t is equally important to detect the emergence of new symptoms (and not only those associated with the initial diagnosis) at any point throughout the treatment course and modify interventions accordingly. At any stage of treatment, but particularly in the acute phase, there may be a dose-response relation between psychotherapy and clinical improvement, and the patient may require more frequent sessions. Short-term augmentation of an antidepressant regimen with a benzodiazepine may assist a depressed and anxious patient, but such treatment should be discontinued as soon as practicable. Initiation side effects must be continually monitored and managed as needed, as these can significantly affect adherence to medication regimens. If pharmacotherapy is used during maintenance and continuation phases, dosing strategies should be based on up-to-date research so as to minimize chances of recurrence. Because there is evidence that psychological therapies can also prolong time to recurrence, at least in depression, such strategies should also be maintained throughout the maintenance and continuation phases.

In any case, clinicians must keep in mind that long-term outcomes are


not especially encouraging. Regardless of the modality used, most patients relapse within a few years of treatment. In one of the longer follow-up studies available, Fava and colleagues (Fava, Grandi, Zielzezny, Rafanelli, & Canestrari, 1996; Fava, Rafanelli, Grandi, Canestrari, & Morphy, 1998) examined outcomes for patients treated with either standard clinical management or CBT. All patients had initially been successfully treated with antidepressant medications. At 4-year follow-up the CBT group had significantly fewer relapses than the clinical management group, but at 6 years there was no significant difference between the two groups, although there were fewer relapses in the CBT group, and that group also experienced fewer multiple relapses.

Maintenance therapy is better than nothing, and more structured maintenance therapies, such as CBT, appear to have some advantage over standard clinical management, as the evidence reported by Fava and colleagues, 1996, 1998) indicates. Just as psychotherapies have relatively poor long-term outcome data, the results for maintenance pharmacotherapy indicate similarly poor long-term outcomes. Nevertheless, evidence suggests that patients who receive adequate pharmacological management experience fewer relapses than those who receive suboptimal management (generally in the form of insufficient dose; it is now well accepted that maintenance doses should be essentially the same as acute-phase doses) or no placebo. For instance, Keller et al. (1998) followed for 19 months patients with either major depressive disorder (MDD) or MDD and dys- thymia who had responded to an acute-phase regimen of sertraline. In this study, 50% of patients treated with placebo experienced new onset of clinically significant depressive symptoms, whereas only 26% of the sertraline-treated group developed such symptoms. Although the data in favor are not overwhelming, Rush and Hollon (1991), after reviewing a number of studies, expressed limited support for the finding that cognitive therapy, whether provided singly or in combination with medication, provided better protection against relapse than pharmacotherapy.

Whatever the choice of intervention, the lowest effective dose should be used. Evidence for this is clearer when pharmacotherapies are used, but a dose-response effect in psychotherapy has been described as well (Barkham, Rees, Stiles, & Shapiro, 1996; Howard, Kopta, Krause, & Or- linsky, 1986). The patient should be continually monitored for the recurrence of symptoms to ensure that the dose is effective, particularly during maintenance and continuation phases. Clinicians must be aware of the pharmacokinetic parameters of drugs used to avoid irrational strategies, such as raising antidepressant doses before an initial response has been determined. These maneuvers do not improve outcome but put the patient a t risk of developing adverse side effects and may contribute unnecessarily to the cost of treatment. The lowest effective dose strategy is particularly germane in the case of the SSRIs, because of the absence of evidence for a dose-response curve with these agents (Gelenberg, 1997). Clinicians should follow the best available evidence in regard to increasing doses or augmenting pharmacological regimens, being aware that such evidence i s being continually updated. Patients should be monitored for the emer-


gence of side effects throughout all phases of treatment, including when pharmacotherapy is discontinued.

When devising a treatment plan, consider the cost of combined or single-modality treatment options. Devising more rational strategies for de- ploying drugs or psychotherapy has important economic ramifications. An- tidepressant drugs are among the top prescribed agents in this country. Fluoxetine was the second-ranking drug in terms of dollar sales in the first 9 months of 1998, with sertraline and paroxetine ranking fifth and seventh. Sixteen percent of all sertraline sales are to elders, and expen- diture for this and other drugs may represent disproportionate demands on the income of those having fixed incomes or inadequate insurance cov- erage (Lagnado, 1998). The introduction of psychotherapy or behavioral management into nursing homes has been shown to reduce the frequency of prescription of psychotropics (Ray et al., 1993; Rovner, Steele, German, Clark, & Folstein, 19921, thereby presumably reducing costs, as well as risks, associated with pharmacological management.

Conclusion

This chapter has taken the reader on a sometimes arduous journey. We have traveled largely in darkness, with precious little in the way of science-based knowledge to illuminate our way. Although there is hope for a more rational future (as demonstrated by the subsequent chapters in this volume), forces other than reason direct much of the clinical practice of pyschopharmacology, and indeed much of all mental health service provision. Healy’s (1997) remark that “There is a real sense at present . . . that knowledge in psychopharmacology doesn’t become knowledge unless it has a certain commercial value. The survival of concepts depends on the interests with which they coincide” (p. 176) can be equally applied to both pharmacological and psychotherapeutic research. The economic and scientific forces that shape psychotherapy research and the provision of psychotherapy are more subtle than those affecting the field of psychopharmacology, but they are by no means absent. In the past, in-depth study of combined treatments for mental disorders has fallen victim to guild- associated biases that have resulted in dichotomous thinking and short- sighted investigative heuristics. This has separated clinical research from much of the reality of everyday clinical practice. Perhaps because the debate has been defined in more global, professional, and academic terms, proponents of either pharmacotherapy or psychotherapy tend to wax mor- alistic about their choices. It is important to resist this temptation. Psy- chotropics have been used extensively throughout the continuum of human history as both intoxicants and therapeutic agents. I t is no more or less moral to seek relief from a pill than it is from a psychotherapist, and moral arguments, although not without inherent seduction, serve little purpose in advancing understanding of the effects of various forms of treatment. We must redefine our interests in terms of our patients. If we are successful in this redefinition, attempts to establish primacy of one


form of treatment over another will fail, and we can direct our energies toward a better understanding of when truly integrated treatments serve our patients best.

References

Ablon, J. S., & Jones, E. E. (1999). Psychotherapy process in the National Institute ofMental Health Treatment of Depression Collaborative Research Program. Journal of Consulting and Clinical Psychology, 67, 64-75.

American Psychiatric Association. (1987). Diagnostic and statistical manual of mental disorders (3rd ed., rev.). Washington, DC: Author.

Appleby, L., Warner, R., Whitton, A., & Faragher, B. (1997). A controlled study of fluoxetine and cognitive-behavioural counselling in the treatment of post-natal depression. British Medical Journal, 314, 932-936.

Barkham, M., Rees, A., Stiles, W. B., Shapiro, D. A., Hardy, G. E., & Reynolds, S. (1996). Dose-effect relations in time-limited psychotherapy for depression. Journal of Consult- ing and Clinical Psychology, 64, 927-935.

Barlow, D. H., & Lehman, C. L. (1996). Advances in the psychosocial treatment of anxiety disorders. Archives of General Psychiatry, 53, 727-735.

Beitman, B. D. (1991). Medications during psychotherapy: Case studies of the reciprocal relationship between psychotherapy process and medication use. In B. D. Beitman & G. L. Klerman (Eds.), Integrating pharrnacotherapy and psychotherapy (pp. 21-44). Washington, DC: American Psychiatric Press.

Bruce, T. J., Spiegel, D. A., & Hegel, M. T. (1999). Cognitive-behavioral therapy helps pre- vent relapse and recurrence of panic disorder following alprazolam discontinuation: A long term followup of the Peoria and Dartmouth studies. Journal of Consulting and Clinical Psychology, 67, 151-156.

Burke, M . J., & Preskorn, S. H. (1995). Short-term treatment of mood disorders with standard antidepressants. In F. E. Bloom & D. J. Kupfer (Eds.), Psychopharmacology: The fourth generation of progress (pp. 1053-1065). New York: Raven.

Buysse, D. J., Reynolds, C. F., Houck, P. R., Perel, J. M., Frank, E., Begley, A. E., Mazumdar, S.. & Kupfer, D. J. (1997). Does lorazepam impair the antidepressant response to nor- triptyline and psychotherapy? Journal of Clinical Psychiatry, 58, 426-432.

Charney, D. S., Berman, R. M., & Miller, H. I,. (1998). Treatment of depression. In A. F. Schatzberg & C. B. Nemeroff (Eds.), Textbook ofpsychopharmacology (2nd ed., pp. 705- 731 1. Washington, DC: American Psychiatric Press.

Chiles, J. A., Carlin, A. S., Benjamin, G. A. H., & Beitman, B. D. (1991). A physician, a nonmedical psychotherapist, and a patient: The pharmacotherapy-psychotherapy triangle. In B. D. Beitman & G. L. Klerman (Eds.), Zntegrating pharmacotherapy and psychotherapy (pp. 105-118). Washington, DC: American Psychiatric Press.

Cooper, P. J., & Murray, L. (1998). Postnatal depression. British Medical Journal, 316, 1884-1886.

Dean, M. (1999). A quiet clinical guideline revolution begins. The Lancet, 353, 651. Depression Guideline Panel. (1993). Depression in primary care: Volume 1. Diagnosis and

detection (Clinical Practice Guideline No. 5 , AHCPR Publication No. 93-0550). Rockville, MD: U.S. Department of Health and Human Services, Public Health Services, Agency for Health Care Policy and Research.

Diket, A. L., & Nolan, T. E. (1997). Anxiety and depression: Diagnosis and treatment during pregnancy. Obstetrics and Cynecology Clinics, 24, 535-558.

Dimeff, L. A., McDavid, J., & Linehan, M. M. (1999). Pharmacotherapy for borderline per- sonality disorder: A review of the literature and recommendations for treatment. Jour- nal of Clinical Psychology in Medical Settings, 6 , 113-138.

Elkin, I . , Shea, M. T., Watkins, J. T., Imber, S. D., Stosky, S. M., Collins, J. F., Glass, D. R., Pilkonis, P. A., Leber, W. R., Doherty, J. P., Fiester, S. J., & Parloff, M. B. (1989). NIMH Treatment of Depression Collaborative Research Program I: General effectiveness of treatments. Archives of General Psychiatry, 46, 971-982.


Fava, G. A., Grandi, S., Zielzezny, M., Rafanelli, C., & Canestrari, R. (1996). Four year outcome for cognitive behavioral treatment of residual symptoms in major depression. American Journal of Psychiatry, 153(7), 945-947.

Fava, G. A., Rafanelli, C., Grandi, S., Canestrari, R., & Morphy, M. (1998). Six year outcome for cognitive behavioral treatment of residual symptoms of major depression. American Journal of Psychiatry, 155, 1443-1445.

Gelenberg, A. J. (1997). Introduction: The practice of pharmacotherapy. In A. J . Gelenberg & E. L. Bassuk (Eds.), The practitioner’s guide to psychoactive drugs (4th ed., pp. 3- 19). New York: Plenum.

Hatsukami, D. K., & Mooney, M. E. (1999). Pharmacological and behavioral strategies for smoking cessation. Journal of Clinical Psychology in Medical Settings, 6 , 11-38.

Haycox, A., Bagust, A., & Walley, T. (1999). Clinical guidelines-The hidden costs. British Medical Journal, 318, 391-393.

Healy, D. (1997). The antidepressant era. Cambridge, MA: Harvard University Press. Hegarty, J. D., Baldessarini, R. J., Tohen, M., Waternaux, C., & Oepen, G. (1994). One

hundred years of schizophrenia: A meta-analysis of the outcome literature. American Journal of Psychiatry, 151, 1409-1415.

Hohagen, F., Winkelmann, G., Rasche-Rauchle, H., Hand, I., Honig, A., Munchau, N., Hiss, H., Geiger-Kabisch, C., Kappler, C., Schramm, P., Rey, E., Aldenhoff, J., & Berger, M. (1998). Combination of behaviour therapy with fluvoxamine in comparison with behaviour therapy and placebo: Results of a multicentre study. British Journal of Psychiatry,

Hollister, L. E., & Claghorn, J. L. (1993). New antidepressants. Annual Review of Phar- macology & Taxicology, 33, 165-177.

Howard, K. I., Kopta, S. M., Krause, M. S., & Orlinsky, D. E. (1986). The dose effect relationship in psychotherapy. American Psychologist, 41, 159- 164.

Hyman, S. E., & Nestler, E. J. (1996). Initiation and adaptation: A paradigm for understanding psychotropic drug action. American Journal of Psychiatry, 153, 151-162.

Jacobson, N. S., Dobson, K. S., Truax, P. A., Addis, M. E., Koerner, K., Gollan, J. K., Gortner, E., & Prince, S. E. (1996). A component analysis of cognitive-behavioral treatment for depression. Journal of Consulting and Clinical Psychology, 64, 295-304.

Jacobson, N. S., & Hollon, S. D. (1996). Cognitive behavior therapy versus pharmacotherapy: Now that the jury’s returned its verdict, it is time to present the rest of the evidence. Journal of Consulting and Clinical Psychology, 64, 295-304.

Janicak, P. G., Davis, J. M., Preskorn, S. H., & Ayd, F. J. (1997). Principles and practice of psychopharmacotherapy (2nd ed.). Baltimore: Williams & Wilkins.

Karasu, T. B., Docherty, J. P., Gelenberg, A., Kupfer, D. J., Merriam, A. E., & Shadoan, R. (1993). Practice guideline for major depressive disorder in adults. American Journal of Psychiatry, 150(4; Suppl.), 1-26.

Keller, M. B., Kocsis, J . H., Thase, M. E., Gelenberg, A. J., Rush, J., Koran, L., Schatzberg, A,, Russel, J., Hirschfeld, R., Klein, D., McCullough, J. P., Fawcett, J. A., Kornstein, S., LaVange, L., & Harrison, W. (1998). Maintenance phase efficacy of sertraline for chronic depression. Journal of the American Medical Association, 280, 1665-1672.

Kirsch, I., & Sapirstein, G. (1998). Listening to Prozac but hearing placebo: A meta-analysis of antidepressant medication. Prevention and Treatment [On-line]. Available: http:// www.apa.org

Klein, D. F. (1998). Listening to meta-analysis but hearing bias. Prevention and Treatment [On-line]. Available: http:// www.apa.org

Klerman, G. L. ( 1991). Ideological conflicts in integrating pharmacotherapy and psychotherapy. In B. D. Beitman & G. L. Klerman (Eds.), Integrating pharmacotherapy and psychotherapy (pp. 3-19). Washington, DC: American Psychiatric Press.

Koren, G., Pastuszak, A., & Ito, S. (1998). Drugs in pregnancy. New England Journal of Medicine, 338, 1128-1137.

Kulin, N. A,, Pastuszak, A,, Sage, S. R., Schick-Boschetto, B., Spivey, G., Feldkamp, M., Ormond, K., Matsui, D., Stein-Schechman, A. K., Cook, L., Brochu, J . , Reider, M., & Koren, G. (1998). Pregnancy outcome following maternal use of the new selective serotonin reuptake inhibitors. Journal of the American Medical Association, 279, 609- 610.

173(S~ppl . 35S), 71-78.


Lagnado, L. (1998, November 17). Drug costs can leave elderly a grim choice: Pills or other needs. Wall Street Journal, pp. 1, 15.

March, J. S., Frances, A., Carpenter, D., & Kahn, D. A. (1998). The Expert Consensus Guide- line Series: Treatment of obsessive compulsive disorder [On-line]. Available: http://www. psychguides.com

Mental health: Does therapy help? (1995, November). Consumer Reports, 743-739. Michels, R. (1999). Are research ethics bad for our mental health? New England Journal of

Medicine, 340, 1427-1430. Miller, I. W., & Keitner, G. I. (1996). Combined medication and psychotherapy in the treat-

ment of chronic mood disorders. Psychiatric Clinics of North America, 19(1), 151-171. Munoz, R. F., Hollon, S. D., McGrath, E., Rehm, L. P., & Vanden Bos, G. R. (1994). On the

AHCPR depression in primary care guidelines: Further considerations for practitioners. American Psychologist, 49, 42-61.

Murphy, G. E., Carney, R. M., Knesevich, M. A., Wetzel, R. D., & Whitworth, P. (1995). Cognitive behavior therapy, relaxation training, and tricyclic antidepressant medication in the treatment of depression. Psychological Reports, 77, 403-420.

Nutting, P. A,, Beasley, J . W., & Werner, J. J. (1999). Practice-based research networks answer primary care questions. Journal of the American Medical Association, 281,686- 688.

Olfson, M., Marcus, S. C., Pincus, H. A., Zito, J. M., Thompson, J . W., & Zarin, D. A. (1998). Antidepressant prescribing practices of outpatient psychiatrists. Archives of General Psychiatry, 55, 310-316.

Pearsall, R., Glick, I. D., Pickar, D., Suppes, T., Tauscher, J . , & Jobson, K. 0. (1998). A new algorithm for treating schizophrenia. Psychopharmacology Bulletin, 34, 349-353.

Persons, J . B., Thase, M. E., & Crits-Christoph, P. (1996). The role of psychotherapy in the treatment of depression: Review of two practice guidelines. Archives of General Psychi- atry, 53, 283-290.

Pincus, H. A., Tanielian, T. L., Marcus, S. C., Olfson, M., Zarin, D. A,, Thompson, J., & Zito, J. M. (1998). Prescribing trends in psychotropic medications: Primary care, psychiatry, and other medical specialties. Journal of the American Medical Association, 279, 526- 53 1.

Pincus, H. A., Zarin, D. A., Tanielian, T. L., Johnson, J. L., West, J . C., Pettit, A. R., Marcus, S. C . , Kessler, R. C., & McIntyre, J . S. (1999). Psychiatric patients and treatments in 1997: Findings from the American Psychiatric Practice Research Network. Archives of General Psychiatry, 56, 441-449.

Ray, W. A,, Taylor, J. A., Meador, K. G., Lichtenstein, M. J., Griffin, M. R., Fought, R.,Adams, M. L., & Blazer, D. G. (1993). Reducing antipsychotic use in nursing homes: A controlled trial of provider education. Archives of Internal Medicine, 153, 713-721.

Redelmeier, D. A,, Rozin, P., & Kahneman, D. (1993). Understanding patients’ decisions: Cognitive and emotional perspectives. Journal of the American Medscal Association, 270, 72-76.

Richards, T. (1998). Partnership with patients: Patients want more than simply information, they need involvement, too. British Medical Journal, 316(7125), 85-86.

Roland, M., & Torgerson, D. J. (1998). Understanding controlled trials: What are pragmatic trials? British Medical Journal, 316, 285.

Rosenheck, R., Tekell, J . , Peters, J., Cramer, J., Fontanan, A,, Xu, W., Thomas, J . , Hender- son, W., & Charney, D. (1998). Does participation in psychosocial treatment augment the benefit of clozapine? Archives of General Psychiatry, 55, 618-625.

Rosser, W. W. (1999). Application of evidence from randomised controlled trials to general practice. The Lancet, 353, 661-664.

Rovner, B. W., Steele, C. D., German, P., Clark, R., & Folstein, M. F. (1992). Psychiatric diagnosis and uncooperative behavior in nursing homes. Journal of Geriatric Psychiatry and Neurology, 5 , 102-105.

Rush, A. J., & Hollon, S. D. (1991). Depression. In B. D. Beitman & G. L. H e r m a n (Eds.), 1n.tegrating pharmacotherapy and psychotherapy (pp. 121-142). Washington, DC: Amer- ican Psychiatric Press.


Sachs, G. S. (1996). Bipolar mood disorder: Practical treatment strategies for acute and maintenance phase treatment. Journal of Clinical Psychopharmacology, 1 G(Suppl.), 32s-40s.

Sammons, M. T., Gorny, S., Zinner, E., & Allen, R. (2000). Prescriptive authority for psychologists: A consensus of support. Professional Psychology: Research and Practice, 31, 604-609.

Schooler, N. J., & Keith, S. J. (1993). The clinical research base for the treatment of schizophrenia. Psychopharmacology Bulletin, 29, 43 1-446.

Seligman, M. E. P. (1995). The effectiveness of psychotherapy: The Consumer Reports study. American Psychologist, 50, 965-974.

Shader, R. I., Fogelman, S. M., & Greenblatt, D. J. (1998). Epiphenomenal, causal, or cor- relational-More on the mechanism(s) of action of antidepressants. Journal of Clinical Psychopharmacology, 18, 265-267.

Shader, R. I., & Greenblatt, D. J. (1993). Use of benzodiazepines in anxiety disorders. New England Journal of Medicine, 328, 1398-1405.

Shekelle, P. G., Woolf, S. H., Eccles, M., & Grimshaw, J. (1999). Developing guidelines. British Medical Journal, 318, 593-596.

Smith, W. T., Londborg, P. D., Glaudin, V., & Painter, J. R. (1998). Short term augmentation of fluoxetine with clonazepam in the treatment of depression: A double blind study. American Journal of Psychiatry, 155, 1339-1345.

Stowe, Z. N., Strader, J. R., & Nemeroff, C. B. (1998). Psychopharmacology during pregnancy and lactation. In A. F. Schatzberg & C. B. Nemeroff (Eds.), Textbook of psychopharmacology (2nd ed., pp. 979-996). Washington, DC: American Psychiatric Press.

Thase, M. E., Greenhouse, J. B., Frank, E., Reynolds, C. F., Pilkonis, P. A,, Hurley, K., Grochocinski, V., & Kupfer, D. J . (1997). Treatment of major depression with psychotherapy or psychotherapy-pharmacotherapy combinations. Archives of General Psychi- atry, 54, 1009-1015.

Thornley, B., & Adams, C. (1998). Content and quality of 2000 controlled trials in schizophrenia over 50 years. British Medical Journal, 317, 1181-1184.

Trivedi, M. H., DeBattista, D., Fawcett, J., Nelson, C., Osser, D. N., Stein, D., & Jobson, K. (1998). Developing treatment algorithms for unipolar depression in cyberspace: Inter- national Psychopharmacology Algorithm Project. Psychopharmacology Bulletin, 34, 355-359.

Valenstein, E. S. (1998). Blaming the brain. New York: Free Press. Wampold, B. E, Mondin, G. W., Moody, M., Stich, F., Benson, K., & Ahn, H. (1997). A meta-

analysis of outcome studies comparing bona fide psychotherapies: Empirically, “all must have prizes.” Psychological Bulletin, 122, 203-215.

Wells, K. B., & Sturm, R. (1996). Informing the policy process: From efficacy to effectiveness data on pharmacotherapy. Journal of Consulting and Clinical Psychology, 64,638-645.

West, J. C., Zarin, D. A,, & Pincus, H. A. (1997). Clinical and psychopharmacologic practice patterns of psychiatrists in routine practice. Psychopharmacology Bulletin, 33, 79-85.

Williams, J . W., Rost, K., Dietrich, A. J., Ciotti, M. C., Zyzanski, S. J., & Cornell, J. (1999). Primary care physicians’ approach to depressive disorders: Effects of physician specialty and practice structure. Archives of Family Medicine, 8 , 58-67.

Woolf, S. H., Grol, R., Hutchinson, A., Eccles, M., & Grimshaw, J. (1999). Potential benefits, limitations, and harms of clinical guidelines. British Medical Journal, 318, 527-530.

Woolf, S. H., & Lawrence, R. S. (1997). Preserving scientific debate and patient choice: Lessons from the Consensus Panel of Mammography Screening. Journal of the Ameri- can Medical Association, 278, 2105-2108.

Zito, J. M., Safer, D. J., dosReis, S., Gardner, J. F., Boles, M., & Lynch, F. (2000). Trends in the prescribing of psychotropic medications to preschoolers. Journal of the American Medical Association, 283, 1025-1030.

1

Combined Treatments for Mental Disorders: Clinical Dilemmas

Morgan ll Samrnons

Believe those who are seeking the truth. Doubt those who find it. -Andre Gide

The absence of a compelling body of evidence on combined pharmacological and nonpharmacological treatments for mental disorders is perhaps the most striking feature of the mental health clinical research literature. This lack of data-particularly in an age of evidence-based practice-about what is arguably the most common form of treatment for mental distress suggests much about the degree to which guild and financial interests shape the pursuit of scientific knowledge. My first task in this chapter is to document the prevalence of combined treatments. I then examine the academic and political phenomena that have contributed to the paucity of data on combined interventions. Obstacles, surmountable or otherwise, to our understanding of these treatments are discussed (along with some occasional successes). I then turn to more practical matters, notably, how one might proceed in developing appropriate standardized protocols that clinicians can use when formulating and applying combined interventions. Because the literature is largely silent, it is difficult to formulate clear, systematic guidelines directing clinicians toward optimum combined treatment strategies. Some tentative guidelines are be offered, but it is ac- knowledged that the current state of understanding renders these guidelines aspirational and, it is hoped, ephemeral, in that directives that are more solidly grounded in science will be forthcoming.

A Failure of Investigative Models: Some Flaws, Fallacies, and Conundrums

Combined drug and nondrug treatments for mental distress are poorly represented in the research and clinical literature. Nevertheless, they are

The opinions expressed by this author represent his views as a private citizen and should not be construed as representing the official opinions or positions of the U.S. Navy or Department of Defense.

11


widespread in clinical practice, to the extent that they may be said to constitute the norm. A significant percentage, possibly even the majority, of all patients receiving services from a psychologist or other nonprescrib- ing mental health practitioner are also simultaneously receiving psychotropic medications, as demonstrated by a number of surveys of mental health service providers (“Mental health,” 1995; Sammons, Gorny, Zinner, & Allen, 2000; Chiles, Carlin, Benjamin, & Beitman, 1991). A further tell- ing indicator of the common nature of combined treatments is the frequency with which primary care practitioners, who are most likely to initially encounter and diagnose mental disorders, use both drugs and referral to mental health specialties. A recent survey demonstrated that 72.5% of depressed patients were given antidepressants, and 38% of these were also referred to a mental health specialist (usually a psychologist or social worker; Williams et al., 1999).

On the other hand, pharmacological treatment has become the mainstay of psychiatric service provision. Reporting on the National Ambula- tory Medical Care Survey data from 1985 and 1993-1994, Olfson et al. (1998) reported that at least one antidepressant was prescribed in 48.6% of all visits to psychiatrists in 1993-1994. Using the same data set, Pincus et al. (1998) discovered that, in 1993-1994, a visit to a psychiatrist specifically for depression resulted in the prescription of a psychotropic agent in 70.9% of cases. Because not all visits to psychiatrists are for depression, the total proportion of visits in which drugs were prescribed was undoubt- edly much higher. This assumption was confirmed by a survey of the practice of 148 psychiatrists in routine outpatient practice (West, Zarin, & Pincus, 1997). In this survey, 90% of all patients of psychiatrists were prescribed at least one psychotropic medication (the mean number of medications per patient was 1.8). In a further analysis of this data set, Pincus et al. (1999) reconfirmed that, in 1997, approximately 90% of patients of psychiatrists surveyed were taking medications. As the authors noted, this was a sizable increase since 1989, when 54.5% of psychiatric patients were prescribed medication. Pincus et al. (1999) also found that 55.4% of out- patients reported on in this survey received both medication and psychotherapy, with psychotherapy being provided either by the psychiatrist or another professional. It is apparent, then, that pharmacotherapy is the mainstay of current psychiatric practice but, even so, the majority of patients also receive psychotherapeutic services. Zit0 and colleagues (2000) also documented a n extraordinary rise in the rate of prescriptions of psychotropics to preschoolers during the 199Os, indicating that the overpre- scription phenomenon is hardly limited to adult populations.

Unfortunately, the pervasiveness of combined treatment is poorly documented in clinical research, and its mechanisms and effectiveness remain the focus of controversy. This in large part may be because of the power of the controlled clinical trial as an investigatory heuristic. Although the benefits of controlled clinical trials cannot be disputed, in certain respects this model has led to an investigative approach that does not capture well the nuances involved in combined treatment. The literature is replete with reports of single-modality, placebo-controlled outcome studies, such as the


effectiveness of cognitive-behavioral models in treating depression. Also, a reasonable number of comparative-treatment outcome studies exist for most major mental disorders. These “horse race” studies often involve head-to-head comparisons of unimodal pharmacological and psychological interventions. Although they have become somewhat less common in recent years (Beitman, 1991), these studies continue to be highly represented in the literature. At the same time, trials of combined treatments are scarce. Only a handful, of variable quality, exist for most disorders.

In part, this situation has been perpetuated by professional biases. Psychologists and nonmedical researchers may have a vested interest in demonstrating the superiority of nonpharmacological techniques. On the opposite side, psychiatric researchers, particularly those with a biological orientation, may tend to champion pharmacotherapeutic strategies. These dichotomous conceptualizations of interventions lead to difficulties in research design and provide a source of investigatory bias that can consid- erably influence outcome. Sources of investigatory bias are difficult to isolate precisely but are reflected by practices such as comparing the treatment being studied against one that appears equivalent but in reality is unequal. One common example of this in drug studies is the strategy of comparing a new drug against an older agent that is effective but has a less favorable side-effect profile. This practice has been found to be ex- tremely common in schizophrenia research (Thornley & Adams, 1998). Re- searchers’ preference for, or allegiance to, one form of treatment over another may also lead to the less favored treatment being inadequately implemented during a clinical trial (Jacobson & Hollon, 1996). A further difficulty in research design is not directly related to hidden researcher bias but is endemic in much of mental health research today. This is the familiar difficulty encountered when efficacy, rather than effectiveness, studies are performed. Efficacy studies, which I discuss in more detail later, comprise the bulk of the scientific knowledge base in mental health research. These studies, usually based on comparisons of two reasonably pure treatments applied in sterile research environments to participants who resemble each other as much as possible, result in outcomes that are poorly generalizable to the everyday treatment setting. As compared to effectiveness studies (examinations of how patients respond to treatments applied in the field; Seligman, 19951, efficacy studies have limited ability to satisfactorily inform clinicians or patients as to optimum choices among treatments (Roland & Torgerson, 1998).

Controlling Bias in Research and Practice

In light of findings that neither psychotherapeutic nor pharmacotherapeutic approaches are superior in the treatment of at least the most common form of mental distress, preference emerges as a key, if not decisive, factor in determining selection of treatment. Preference may be expressed by either the clinician or the patient. Patient choice is an important variable in determining positive outcome, but patient preferences are probably in-